Effects of Pilot Programs for China’s Central SOE Board of Directors

LiWengui(李文贵),YuMinggui(余明桂)andZhongHuijie(钟慧洁)

Abstract: Since June 2004, the State-Owned Assets Supervision & Administration Commission (SASAC) has launched pilot programs for a board of directors (BoD) for central SOEs to establish and improve their governance structure and standardize their exercise of shareholder rights over state-owned listed companies. Based on this quasinatural experiment, this paper examines the BoD’s effects on the agency cost of state-owned listed companies and their economic consequences. Using data of central SOE-controlled companies listed at Shanghai and Shenzhen stock exchanges during 2002-2015, this paper finds that the pilot programs significantly reduced the two types of agency costs for the companies, and such effects primarily existed for listed companies with smaller central SOE shareholding ratios. Further test uncovers that, compared with central SOEs that did not carry out the pilot programs, those that did reported higher economic value-added and stock returns. Our conclusions offer a new interpretation of the BoD’s governance effects from a controlling shareholder’s perspective, and provide empirical evidence for the positive effects of the pilot programs for central SOE boards of directors. These findings have important policy implications for deepening SOE governance reforms.

Keywords: central SOE, board of directors, agency cost, shareholding ratio, corporate performance

JEL Classification Codes: G32; G38; P31

DOI:1 0.19602/j .chinaeconomist.2018.11.081 0.19602/j .chinaeconomist.2018.09.02

1. Introduction

The Chinese government has carried out various reforms, including ownership and governance reforms, to address SOE inefficiency. For instance, the State-Owned Assets Supervision & Administration Commission (SASAC) introduced pilot programs to create a board of directors for central SOEs - SOEs directly affiliated to SASAC - in June 2004 in a bid to improve corporate governance structure. By the end of 2015, a total of 85 central SOEs were included in the scope of such pilot programs1. During

this period of time, effects of these pilot programs were widely discussed. Some scholars believe that the internal governance mechanism and strategic management of central SOEs substantially improved after implementation of the pilot programs (Qin, 2012; Gao, 2012). Others contend that many external directors in the pilot programs lacked real power, that the BoD’s functions did not fully materialize, and governance issues were left unaddressed (Xiong, 2015; Zhou, 2016). Regretfully, neither contention came from strict theoretical and empirical studies and was more than just points of view. As such, this paper intends to analyze and validate the economic consequences of the pilot programs for central SOE board of directors from the perspective of the agency problems of SOE-controlled listed companies.

Boards of directors ( BoD) and relevant institutional arrangements are core mechanisms for mitigating firms’ agency problems (Fama and Jensen, 1983; Cai et al., 2015). From the BoD’s decisionmaking and supervision functions, the literature analyzed how the BoD’s independence, structure, size, efficiency, culture and other attributes influence listed companies’ agency problems, as well as economic results. Their findings uncover that the BoD is conducive to the governance effects of shareholder supervision and executive pay incentives (Zheng and Lü, 2009; Balsmeier et al., 2017) and is significantly positively correlated with firm efficiency (Wang et al., 2006; An et al., 2016). The BoD’s participation in strategic decision-making can improve a firm’s decision-making quality (Wang et al., 2006). However, different board compositions and structures exert heterogeneous effects on firms’ strategic options and decision-making efficiency (Xie and Zhao, 2011; Xie et al., 2011). The BoD size and frequency of board meetings may also influence firms’ agency problems and reduce corporate financial fraud to some extent (Yang, 2009). Zheng (2012) finds that good board culture prevents excessive executive compensation.

These studies provide empirical evidence for board governance, but they all merely directly test the BoD’s corporate effects on listed companies. Based on the pilot programs for central SOE board of directors as a quasi-natural experiment, this paper investigates the BoDs’ effects on state-owned listed companies’ agency problems and performance. Based on the data of SOE- controlled listed companies during 2002-2015, this paper finds, through a difference-in-differences (DID) model, that after implementation of the pilot programs, the overhead expenses and other receivables of treatmentgroup enterprises reduced more significantly compared with control-group enterprises. The implication is that the pilot programs mitigated state-owned listed companies’ agency problems. Equity structure is an important factor that influences the controlling shareholder’s behaviors. As a result of the higher shareholding ratio, the controlling shareholder shares interests consistent with those of the listed company. This makes the controlling shareholder more motivated and capable of supervising the listed company and more likely to refrain from entrenchment (Claessens et al., 2002). After grouping listed companies by SOE shareholding ratio, we find that the pilot programs significantly mitigated agency problems only for listed companies with smaller SOE shareholding ratios. Further test discovers that, compared with the control group that did not implement the pilot programs, treatment-group firms reported significantly higher economic value-added and stock return after the creation of board of directors for the SOE.

This paper may offer the following contributions:

First, it provides empirical evidence on the positive effects of the pilot programs for central SOE board of directors. Since being launched by the SASAC in June 2004, the pilot programs aroused continuous discussions on their effects, including positive effects and inadequacies. But this topic is yet to be discussed in theoretical studies. Using a difference-in-differences (DID) model, this paper reveals the pilot programs’ effects on listed companies’ agency problems and economic consequences, uncovering a significant fall in pilot listed companies’ agency costs and a major improvement in their performance. From a theoretical view, this paper offers new empirical evidence on how the pilot programs mitigated listed companies’ agency problems and improved their performance.

Second, this paper offers new interpretations on the BoD’s governance effect. According to the test

result of the BoD’s independence, structure and culture in existing literature, the BoD’s increasingly sophisticated system shows a significant governance effect in the Chinese market (Yang, 2009; Zheng and Lyu, 2009; Xie, et al., 2011; Zheng, et al., 2012; et al.). However, Cai et al. (2015) finds in case studies involving Suntech Power that even a BoD system formed according to strict regulation rules did not prevent some firms from failing, and thus puts forward the “board of directors puzzle.” Unlike the above studies that directly analyze listed companies’ BoD, this paper aims to uncover the BoD’s effects on China’s state-owned listed companies based on the pilot programs for central SOE board of directors. From a controlling shareholder’s perspective, this study further enriches and expands research on the BoD’s governance effects.

Third, this study offers policy implications for further reforming central SOE governance. The pilot programs are an important part of China’s overall SOE reform. The intention is to standardize central SOEs’ exercise of shareholder’s rights by improving internal governance, so as to achieve listed companies’ better performance. Since Baosteel became China’s first pilot enterprise, controversies over the pilot programs’ effects have persisted. We notice that the pilot programs mitigated SOE-controlled listed companies’ agency problems (especially those with smaller shareholding ratios) and significantly increased their economic value-added and stock return of listed companies. By verifying the pilot programs’ positive effects, this paper provides a theoretical basis for further development of SOE board of directors.

2. Institutional Background and Theoretical Analysis 2.1 Institutional Background

As the controlling shareholder of many state-owned listed companies, most solely state-owned enterprises (SOEs) in China were created under China’s Company Law of 1988 without a board of directors. Even for a minority of such SOEs with a BoD, many of their board members concurrently assume management positions. Without separating decision-making powers from executive powers, the BoD’s roles cannot be brought into full play (Xu, 2011). In February 2004, SASAC proposed to the State Council to initiate the pilot reform of the SOE board of directors. The reform aims to create and improve solely state-owned companies’ governance structure and enhance their performance. In June 2004, SASAC released the Notice on the Creation and Improvement of BoD Pilot Programs for Solely StateOwned Companies. In March 2009, SASAC released the Interim Measures for the Standard Operation of Boards of Directors of Central SOEs under the Pilot Programs to further implement the pilot reform.

The reform aims to devolve government powers, i.e. central SOEs should first create a standard board of directors, and SASAC will devolve powers (e.g. powers for decision-making of mid -and long-term development, management personnel’s appointment and performance evaluation, and major financial matters) to the board of directors (Qin, 2012). A standard board of directors should (1) create a system of external directors for the latter to supervise the management and take part in making investment and financing decisions and daily operations; (2) create ad hoc committees to ensure the board’s effective strategic control and supervision (Xu, 2011); and (3) implement a board reporting system. By June 2017, almost 90 central SOEs were incorporated into the scope of the pilot reform, and 83 of them created a standard board of directors2. For the central SOEs involved in the pilot programs, their external directors accounted for over 50% of board membership. Most of them created ad hoc committees for compensation, audit, strategic planning and nomination, and submitted annual work reports of the board to SASAC on a regular basis.

2.2 Theoretical Analysis

2.2.1 Pilot programs for central SOE board of directors and agency cost of state-owned listed companies

According to agency theory, the controlling shareholder has an important influence on companies’ agency problems (La Porta et al., 1999). The controlling shareholder is motivated and able to supervise corporate operations, and prevent them from seeking self-interest, thus mitigating the first type of shareholder-management agency problems (Grossman and Hart, 1988), i.e. “supervision effects.” On the other hand, the controlling shareholder may also cause the second type of agency problems between the controlling shareholder and other shareholders (La Porta et al., 1999), i.e. “entrenchment effects” by manipulating corporate decision-making and encroaching upon corporate interests.

As controlling shareholders of many state-owned listed companies, central SOEs are hampered by their governance problems in performing shareholder’s rights (Lin, 2012). In terms of supervision effects, central SOEs are agents in the multi-tiered agency relationship of state ownership - such an agent identity is insufficient to motivate them to supervise state- owned listed companies ( Zhang, 1999). Regarding entrenchment effects, as solely state- owned entities, central SOEs do not have shareholders’ meetings, and their board members also assume management positions. Some noncorporate central SOEs do not even have a board of directors. Such SOEs are influenced by government in making decisions and performing their shareholder’s rights over listed companies. The government may intervene to achieve political objectives, thus distorting the central SOEs’ business objectives and causing them to encroach upon listed companies’ interests (Wu, 2001). Central SOE leaders are nominated by the government and are not subject to effective supervision, thus causing serious insider control problems (Qin, 2009). Such insider control creates opportunities and motivations for both central SOE executives and the government to encroach upon listed companies’ interests.

The pilot programs for the SOE board of directors may increase supervision over state-owned listed companies, and diminish central SOEs’ entrenchment effects on state-owned listed companies. First, creation of a standard board of directors will improve central SOEs’ governance structure. This helps mitigate central SOEs’ agency problems, and motivate them to supervise the listed companies. Second, government institutions like SASAC will devolve capital contributors’ powers to the board through the pilot reforms, including powers for decision-making on major investment and financing activities, internal reform and reorganization, shareholders’ power for state-owned firms and appointment of management personnel (Qin, 2009; Xu, 2011). To some extent, this will reduce government intervention in central SOEs and state-owned listed companies. It will also increase the cost for government or officials to intervene in business operation. Third, a standard board of directors requires external directors who account for more than half of the central SOE board members, as well as ad hoc committees where external directors participate in corporate decision-making and daily operations. External directors are appointed and primarily paid by SASAC. They are independent from central SOEs and participate in decision-making based on the goal of maximizing corporate value3. Thus, central SOEs’ decisionmaking and executive powers will be separated to mitigate insider control problems. Board supervision and particularly external directors’ supervision on SOE management will curb self-interested behaviors. With science-based decision-making and check and balance that lead to diminished intervention and

insider control, central SOEs will refrain from encroaching upon state-owned listed companies’ interests.

Hypothesis 1: The pilot programs for central SOE board of directors will significantly reduce SOEcontrolled listed companies’ agency cost.

2.2.2 Pilot programs for central SOE board of directors, equity structure and agency cost of SOEcontrolled listed companies

A central SOE may have controlling shares in multiple listed companies. The question is whether the pilot programs for a board of directors have differentiated effects on different listed companies? According to agency theory, equity structure is an important factor that influences the controlling shareholder’s “supervision effects” and “entrenchment effects.” Therefore, the pilot programs’ agency cost effects may be subject to equity structure.

A more concentrated equity structure has important governance effects ( Shleifer and Vishny, 1997). Higher shareholding ratio motivates the controlling shareholder to supervise a listed company’s management personnel ( Chen et al., 2015). Of course, such supervision has certain costs. If the controlling shareholder’s shareholding ratio is too low, smaller shareholders will benefit from the company’s increased value from supervision through free-riding. As a result, the controlling shareholder assumes an additional cost of supervision, and benefits from a less marginal return of supervision (Jameson et al., 2014). For instance, Wang and Zhou (2006) find that a higher ownership ratio creates a stronger incentive for the controlling shareholder. Du et al. (2016) uncovers that the higher shareholding ratio of the controlling shareholder is conducive to restraining excessive investment. Kang et al. (2017) provides evidence that the controlling shareholder’s high shareholding ratio is conducive to a company’s longterm value.

On the other hand, a concentrated equity structure is believed to cause more serious agency problems for the controlling shareholder (Johnson et al., 1996). Higher shareholding ratio will create opportunities for the controlling shareholder to exploit a listed company’s resources. For instance, a controlling shareholder may encroach upon a listed company’s interests through related transactions, assets reorganization, capital occupancy or dividend distribution policy (Li, 2005; Wang, 2007; Jiang et al., 2010; Jameson, 2014). The controlling shareholder may benefit from shared income and private income (Grossman and Hart, 1988). When encroaching upon a listed company’s interests, the controlling shareholder not only obtains private income but also reduces the company’s value and shared income as well. When the controlling shareholder’s shareholding ratio is too high, such an act of infringement will harm shared income even more. Tu and Liu (2010) discover that with an increasing shareholding ratio, a listed company’s controlling shareholder becomes more able but less motivated to exploit its resources. As a result, the actual ratio of funds occupied by the controlling shareholder is smaller.

If a central SOE holds a higher ratio of a state-owned listed company’s shares, it may theoretically increase supervision over the listed company, and become less motivated by political goals or selfinterest to exploit the company’s resources. This is true for both government entities and central SOE executives. Therefore, the pilot reforms for central SOE board of directors may play a smaller role in mitigating the agency problems for listed companies with more concentrated equity structure.

Hypothesis 2: Compared with listed companies with higher shareholding ratios of controlling shareholders, the pilot programs for central SOE board of directors exert a greater influence on those with smaller shareholding ratios.

3. Research Design 3.1 Sample and Data

The research period of this paper is 2002-2015. According to report of Economic Daily and other sources, 85 SOEs were brought into the scope of the pilot programs for central SOE board of directors.

Based on the official websites of SASAC and central SOEs, as well as search engines like Baidu.com, we compile a list of 71 central SOEs under the pilot programs with specific years of implementation. Time points include: first, the time of inclusion of SOEs into the pilot programs as determined by SASAC; second, the time of creation of the board of directors by these SOEs. For instance, seven SOEs including Baosteel Group and Shenhua Group were identified as the first round of the pilot programs in June 2004. However, the two did not hold formal board working sessions until October 17, 2005 and November 25, 2005, respectively. This paper considers the “convening of board working sessions” as the event time of the pilot programs. Then, based on the official websites of the central SOEs, we collect 244 sample companies listed at Shanghai and Shenzhen stock exchanges.

Other data employed in this paper are from the Wind and CSMAR databases. Upon validation, we delete central SOEs involved in consolidation during the sample period, as well as their controlled listed companies, and delete samples with asset-to-liabilities ratio greater than 1. Lastly, after excluding samples with missing financial data, we obtain 2,845 valid observations.

3.2 Model Specification and Variable Definition

This paper adopts a difference- in- differences ( DID) model to analyze the effects of the pilot programs for central SOE board of directors on the agency cost of state-controlled listed companies. The model to be validated is specified as follows:

(1) In Model (1), Agency is the agency cost of firms. According to Ang et al. (2000) and Li and Yu (2015), the overdead expense ratio ( Agency1) is selected to measure the first type of agency cost. Li (2007) believes that funds of listed companies directly occupied by the majority shareholder are often included into other receivables. Thus, other receivables ( Agency2) is used to measure the second type of agency cost, which is defined as other receivables of the company by the end of period divided by total assets. Board is the dummy variable of sample attribute. Based on data after the implementation of the pilot programs, if sample companies’ controlling shareholders had implemented the pilot programs during 2005-2012, the companies are classified as “treatment group,” and the value of Board is 1. If sample companies’ controlling shareholders did not implement the pilot programs prior to 2015, such companies are classified as “control group,” and the value of Board is 0. After is the dummy variable that denotes time attribute. For years prior to the implementation of the pilot programs, the value is 0, and during the implementation of the pilot programs and thereafter, the value is 1.

X denotes a series of control variables, including: (1) company size ( Size), defined as the natural logarithm of a company’s total assets at the end of the period; ( 2) asset- to- liabilities ratio ( Lev), defined as total liabilities of a company at the end of the period divided by its total assets; (3) corporate accounting performance ( Roa), defined as a company’s net profits divided by its total assets; (4) age of listing ( Age), defined as Ln (age of the company’s listing+1); (5) slales growth rate ( Growth); (6) cash holdings ( Cash), defined as the balance of a company’s cash and cash equivalents at the end of the period divided by its total assets; and (7) the largest shareholder’s shareholding ratio ( Shareholder1). In order to control for the board governance effects of listed companies themselves, we also include two variables of the proportion of independent directors of a listed company ( Rindep) and whether the four committees ( Committee) are created.

3.3 Descriptive Statistics

Table 1 reports major variables’ descriptive statistical characteristics. In order to remove outliers, we conduct Winsorize treatment at 1% for continuous variables that reflect financial characteristics. Average ratio of overhead expenses to a company’s total assets is 3.99%, and average ratio of other receivables to a company’s total assets is 1.54%. According to Jiang and Yue (2005), the average ratio of

other receivables to Chinese listed companies’ total assets was 9% during 1996-2005 - an earlier period preceding our study’s timeframe, which is much higher compared with samples in this paper. This may suggest that, with China’s deepening market-based reforms and increasing regulatory intensity, there were some mitigations in the agency problems where majority shareholder occupied listed companies’ funds.

4. Analysis of Empirical Result 4.1 Pilot Programs for Central SOE BoD and Agency Cost of State-Owned Listed Companies

Table 2 reports the test result of the impact of the pilot programs on state-owned listed companies’ agency cost. From columns (1) to (6), the coefficients of interaction term Board ×After are significantly negative at 5% and 10% levels respectively. The implication is that, compared with sample SOEs that did not implement the pilot programs, those that did reduced much more overhead expenses and other receivables within three years after implementation. This result suggests that after central SOEs created BoD and introduced supervisory mechanisms like external directors, the listed companies in which they had controlling shares experienced mitigations for both types of agency problems. This finding offers empirical evidence for the positive effects of the pilot programs for central SOE board of directors.

4.2 Pilot Programs for Central SOE BoD, Equity Structure and Agency Cost of SOE-Controlled Listed Companies

By central SOEs’ median shareholding ratios for listed companies in each year, we divide total samples into two subgroups with high and low shareholding ratios respectively, so as to reveal whether the pilot programs’ positive effects on listed companies’ agency problems were influenced by their equity

structure.

Table 3 shows that the coefficients of interaction term Board ×After are not significant in Columns (1) and (3), but significantly negative at 5% level in Columns (2) and (4). This implies that the pilot programs primarily influenced the agency cost of listed companies in which central SOEs had smaller shareholding ratios. This finding supports Hypothesis 2. We believe that when their shareholding ratio is

high, SOEs supervise listed companies more effectively, and refrain from intervention or insider control that lead to entrenchment behaviors. Thus, the pilot programs did not significantly mitigate agency problems for listed companies in which SOEs held high stakes, and the effects are more significant for sample group with a smaller SOE shareholding ratio.

A possible implication is that reducing state ownership through share transfer, i.e. privatization, is

not the only option to enhance SOE performance. For one thing, appropriate ownership concentration may enhance governance effects. For another, without changing controlling shareholders’ shareholding ratios, state-owned listed companies may perform better if we standardize controlling shareholder’s exercise of shareholder’s rights to mitigate their agency problems. This finding is consistent with Yu’s et al. (2016) view. Based on central SOE executives’ performance evaluation indicators, Yu et al. (2016) argues that reducing state ownership, i.e. privatization, may not be the only way to address SOE inefficiency.

4.3 Robustness Test

4.3.1 Analysis of the financial characteristics of different sample groups prior to the pilot programs

The SASAC and other government departments had a major influence over whether central SOEs carried out the pilot programs for board of directors. Therefore, we conduct a further comparative analysis of the financial characteristics of control group and treatment group prior to the implementation of the pilot programs to shed light on the possible existence of sample selection bias between different test groups. Data in Table 4 show that the two sample groups have no significant differences in terms of company size, growth opportunities and cash positions no matter by mean value or medium value. As for asset-to-liabilities ratio, the difference of their mean values is not significant, and their medium values exhibit differences at about 10% level. In terms of the age of listing, control group’s value is significantly below that of treatment group at 5% level no matter for mean value or median value. The implication is

that for listed companies with a longer history of listing, their shareholding central SOEs are more likely to be brought into the scope of pilot programs for board of directors. On the whole, only one of the five financial indicators of control group and treatment group has significant difference, which indicates that no serious problem of sample self-selection exists between control group and treatment group4.

4.3.2 Selection of paired samples

According to Table 4’s analysis result, there are some differences between control-group sample firms and treatment-group sample firms in terms of the age of listing and asset-to-liabilities ratio. Hence, we select paired samples for treatment group by asset-to-liabilities ratio and age of listing. Test result is shown in Table 5. Coefficients of interaction term Board×After in Columns (1) and (4) are significantly negative at 5% and 10% levels respectively. This result is consistent with Table 2, and indicates that the

pilot programs for central SOE board of directors significantly reduced SOE-controlled listed companies’ G&A expense ratio and other receivables, thus mitigating their agency problems. Although coefficients of interaction term Board ×After in Columns (2) and (5) are all negative, none of them is significant. Coefficients of interaction term Board ×After in Columns (3) and (6) remain significantly negative at about 5% level. This implies that the pilot programs mainly influenced listed companies in which central SOEs had smaller shareholding ratios, but did not significantly mitigate the agency problems for those with higher SOE shareholding ratios.

4.3.3 Other robustness test

(1) Observations of all sample firms in the year of the pilot programs are deleted to directly test the status of sample firms three years before and after the pilot programs. (2) Sample firms that carried out the pilot programs in 2012 are deleted, and “treatment group” is defined as listed companies controlled by central SOEs that implemented the pilot programs during 2005-2011. (3) Net amounts of capital investment and capital occupancy are used as indicators to measure both types of agency cost. Capital investment (Invest) is defined as (capital expenditure + M&A expenditure - income from sales of long-

term assets - depreciation) / total assets. Capital investment usually involves riskier investments with longer cycles. In case of serious agency problems, listed companies’ managers will invest less capital to keep an easier workload or due to career considerations (John et al., 2008). According to Li and Zheng (2015), the net amount of capital occupancy refers to the net amount of other receivables minus the listed company’s other payables that occupy the controlling shareholder’s capital. Relevant regression result is reported in the first two columns of Table 6. Similarly, the coefficients of interaction term Board×After are always significant at 5% or 1% level.

5. Further Validation

Since the pilot programs for central SOE board of directors will mitigate the agency problems of SOE-controlled listed companies, the question is whether these companies’ performance will improve as a result of such pilot programs. In fact, a primary objective of these pilot programs was to improve state-owned listed companies’ performance by standardizing central SOEs’ exercise of controlling shareholder’s rights. Thus, we further analyze the effects of the pilot programs on listed companies’ performance to gain a better understanding of their economic consequences.

We use economic value added (EVA) and annual return to company stocks (Return) to measure state-owned listed companies’ performance. Explained variable of Columns (1) and (3) in Table 7 is EVA per share, and the coefficients of interaction term Board×After are always significantly positive at 1% level. This implies that, compared with SOE-controlled listed companies that did not carry out the pilot programs, sample firms that did reported higher EVA per share within three years after implementing

the pilot programs. Explained variable in Columns (2) and (4) is return to individual stocks, and the coefficients of interaction term Board×After are significantly positive at 1% and 5% levels respectively. This implies that the SOE-controlled listed companies reported much higher stock returns within three years after the pilot programs. That is to say, the pilot programs for SOE board of directors led to much better market performance, and brought higher returns to shareholders. The market responded positively to the pilot reforms for central SOE board of directors.

6. Concluding Remarks

This paper investigates how SASAC’s pilot programs to create boards of directors ( BoD) for central SOEs influenced state-owned listed companies’ agency costs and economic results. We collect information about the creation of a board of directors by central SOEs during 2000-2015, and classify SOEs that implemented the pilot programs as treatment group and those that did not as control group. Using a difference-in-differences (DID) model, we discover that after the pilot programs, treatmentgroup firms’ overhead expenses and other receivables reduced by a much greater margin compared with control-group firms. This implies that the pilot programs mitigated state-owned listed companies’ agency problems. As can be found from validation after grouping listed companies by SOE shareholding ratios, the pilot programs for a board of directors for SOEs significantly mitigated the agency problems only for the sample group with smaller SOE shareholding ratios, and the effects were insignificant for the sample group with higher SOE shareholding ratios. Further test reveals that treatment-group firms reported significantly higher economic value-added and stock returns after the creation of a board of directors for the SOEs. The above results provide empirical evidence for the pilot programs’ positive effects from an agency theory’s perspective, and help clarify controversies surrounding the effects of the pilot programs. They also provide a new interpretation on BoD’s governance effects from a controlling shareholder’s perspective. For these reasons, this paper’s findings are of important policy implications for further deepening the governance reform of China’s central SOEs.